Reducing Cloud Computing Costs with Workspot’s Brad Tompkins

This is Textron tv. Hey guys, thanks for the throw. We're here with Brad Tompkins, who's CEO for Work Spot, and we're talking about how odd it is that cloud computing pricing seems to be consistent everywhere you go, and we're not quite clear what's going on with all of that.

Brad, welcome to the show. Thank you. Thanks for having me, Mike.

Great to be here. So is this thing with cloud pricing that's so remarkably consistent, a happy accident? Or is something else at work here?

Well, it's hard to say. Um, you know, cloud pricing is, has always been kind of high. Uh, when you put everything together, uh, on the surface, it it seems like it's, uh, you know, a good deal In some ways when you look at total cost of ownership, uh, pricing is actually better than, than running your own data centers, uh, and easier and more convenient.

Um, but from a, a vendor perspective and, and the prices they're charging, you know, I'm, I'm sure they put quite a bit of thought into, into how this goes, uh, out to the public. But, uh, it does seem to be pretty consistent and, uh, it is, it is very interesting and, and the impacts it's having on the market and on, uh, IT budgets And for a lot of services, at least the mainstream ones, we don't see a lot of movement in the pricing. It seems like, you know, there'll be drops in prices on, uh, ancillary services all the time, but I don't think there's a whole lot of drops in pricing when it relates to, um, the basic services that everybody seems to use are the same almost, you know, month in and month out.

Yeah. You know, I mean, the reality is is that, you know, they have to make their profits and, and they have their business plans and their profit margins that, that they need to make. Um, you would think that, uh, with more volume comes a, a decrease in pricing.

But then again, actually to be fair, if you look at the increased costs on the infrastructure side from, uh, the likes of Nvidia, uh, and for customers that are using expensive GPUs, those prices are probably gonna go up, not down. Right. And, uh, so I think everybody has to take a look at their own sit, uh, individual situation.

Um, and I do know that there are some providers out there that are looking, you know, to cut costs where they can. Um, and hopefully that, uh, that translates to lower prices to, to customers. Now there are other things that customers can do in IT departments can do to help keep their prices down.

And two things like hibernate and start, stop, and of course all those things are built into work spot for, for our solutions. But, uh, there are things that, that people can do to keep costs down. It almost seems like it's difficult for people to master all of that.

'cause it seems like there are, uh, reserve instances and spot instances, but it's not clear to me people are using those aggressively. And this a tendency to maybe just to default to a standard usage. 'cause I don't want to think too hard about this stuff.

So, um, how do we make that easier? 'cause it seems highly dynamic and highly complex. You know, it isn't easy and you do have to keep your eye on the ball.

Uh, there are some automated functionalities and features and products that are out there. We have some built into our products that are automated. Um, but you do have to set things up properly with the provider to make sure that you can, can take advantage of those things.

There are other third party products out there from IBM and other large providers that can help people keep, you know, tabs on their usage. Uh, I know that storage overruns is a problem, right? And there are ways to, to stay on top of that, it's not easy.

Um, you know, we do a lot of business in the cloud and, and even sometimes, you know, we find that we miss a thing or two and we have to go back and and remediate on our own. But it, it's, it is a, it is a full-time job to make sure that you're, you're utilizing all those features that can help you save money. We have seen the rise of finops teams.

Is that the full-time job that you kind of described and what's driving that? And I'm kind of curious whether you see that as a long-term phenomenon, or is it a short-term thing while we figure out how to master it? It's probably long-term from my point of view.

I mean, the cloud's not going anywhere. Um, it's, there's just, there's too much convenience there. And you know, we have our own FinTech folks that, that keep an eye on our spend.

And I, I don't know that it'll go anywhere until it gets fully automated. Right. And when they're fully automated products and perhaps AI is the answer, um, or machine learning as a part of AI to, to help us understand, you know, where those cost overruns are happening.

You know, I think that, that people are gonna have to keep an eye on it for a while. Well, I'm asking that question 'cause it historically, at least we gave developers a credit card and they would off and provision these things, but I don't think they really understood the costs involved in deploying those workloads. So, um, where does that adult supervision finally come into play?

Well, it has to be through process as much as anything, right? And yes, there was a time when, you know, you took a credit card and you swiped the credit card and you got some, some cloud usage. And then some people got surprised and they got a bill that the number at the bottom looked more like a phone number than, than an actual bill.

Uh, and so the, the really, the way around this is as, as an IT department and as a company, you have to put a process in place to keep this from happening. You'll never stop all of shadow it, if you want to think of it that way. Um, but you put the right process and implement the process, you can get your arms around it.

And the other thing to do too is by the way, is, you know, put together the right commits with those cloud providers so that you have plenty of room to do what you need to do. Um, but obviously don't overcommit. 'cause then that puts you in a different situation.

But, uh, there are, there are things you can do and most of them are process oriented, I would argue. We've also seen a lot of organizations are signing these enterprise agreements. Um, but it's not clear to me that that's the most efficient way to go.

'cause you hear from some of them and they have, uh, at the end of the year, they have capacity they never used. So they overpaid for it. Um, yeah.

How, how challenging is it to manage these annual agreements that are out there? You know, enterprise agreements or ELAs as they're sometimes called are shared risk, right? So the, the customer takes some risk that they may not utilize, uh, everything they paid for.

And then the vendor takes some risk that you may over utilize and, and they may have discounted more than they, than they thought they wanted to, right? So generally, not all the time, but a lot of these ELAs are shared risk scenarios. And I think, you know, one has to really put a lot of thought into what they're gonna commit to before they get into them at the end of the day.

Um, so it could go either way. Uh, as, as a, as a provider of ELAs not at work spot, but at previous companies we, I provided ELAs where we were, we were not the beneficiary of the good part of the deal. Right?

Um, so it, it can, the shared risk can go either way. We have seen all these workloads get deployed, but I don't often see a workload move from one cloud to another cloud. It seems people kind of keep those there.

It's a heavy lift. So do they really have any kind of, uh, redress here if they decide that something's too expensive? I mean, or am I just basically once I make that initial decision, I'm kind of in for the run?

I think it's more about the more about the commits that you have with that provider than it is about the technology to, to move workloads. We move workloads all the time between, 'cause we're sort of neutral the way we think about cloud providers with, uh, VDI workloads that we service and we can move them around. Um, it takes a little bit of work, but it can definitely be done.

And so I don't, I don't see like a suit there. There's a little bit of lock in, especially if you do a big commit. If you do a big commit, there's a lock in, right?

Um, technology wise, it's not as sticky as, say, trying to change out an ERP system, for example. So I guess it's all relative. Do you think there might come a day when maybe there's a congressional committee that gets together and starts asking some questions about, you know, how all this pricing comes together?

And it seems like there might be a call for maybe a little more transparency Potentially? Um, I, I do also see the challenge maybe a little bit with pricing, right? I doubt and I don't wanna say there's any collusion going on 'cause I just, I personally don't see that.

Right. Um, but you know, there are other issues in terms of licensing, right? So, uh, can I use X vendor's product on the other vendor's cloud?

Right? I think that's a problem that needs to be addressed, um, maybe even more so than the pricing, because that does create a lock in situation. Hmm.

How savvy do you think folks are about the cloud today? I mean, um, we've been at it for 10 years, but I, I have to wonder, I mean, some organizations, I think if you're, I don't know, uh, fortune 50, you probably got the resources to become an expert in this space. But the average IT organization, do they kind of get it?

Or are they still just kinda, uh, haphazardly navigating their way through from workload to workload? I think, I think they could use some help. The smaller organizations, you know, there are great partners out there that can help great cloud partners that, that can help with this.

You know, the adas of the world and the software ones of the world, you know, these, these guys were born from the, from the cloud and they can, there, there are plenty of consultants out there that can help smaller businesses navigate this and get their arms around costs and how they want to use it. And there's a good market for that out there. You mentioned ai.

Do you think we'll get to the point someday when I'll just come into the office and there'll be a danger will Robinson message from some AI bot somewhere saying, uh, you know, this workload is about to exceed its pricing cost thresholds and I should do something about it? For sure. That, that's gonna happen for sure.

Uh, and I don't mean like the dystopian version of what you said. I mean the, the good version of what you said, the, uh, the, uh, you know, hey, you're about to overrun some costs, let's move some workloads around or shut some of these things down. We're actually working on some of this AI capability in our product, uh, for, uh, virtualization workloads and BDI workloads, um, as well as, uh, understanding observability and what may be going wrong.

Right? Um, and you could have something go wrong where you have a a, a process that's spinning outta control. And yes, we can see it now and we can alert the person now without ai, but when we bring AI into the mix, we can maybe even help remediate it or see it before it's even gonna happen because we can see patterns across our customers where the AI can see it.

Right? So this is absolutely coming in the good way. The the dystopian way is a whole nother conversation maybe for another interview.

There you go. Um, why are we having this conversation about cost now? I mean, we've been at this for a while.

I mean, did we all wake up one morning and go, Hey, this is suddenly expensive, or did something change in the landscape? That, that's a really, really good question. Um, you know, I have seen some data, some feedback from some of the analysts that there's actually a repatriation of some workloads coming back on premises for some of the larger organizations, especially because they have good controls and good processes to run a good data center.

They know what they're doing. Um, and I think cost is a big part of this. Um, I also think that having control is another part of it.

And, and the reality is is clouds aren't, they're not perfect. They can and have gone down, right? Um, and if that happens and it's in your large organization with regulatory compliance requirements and that cloud went down, you can't point to the cloud provider and say, oh, well that's their fault.

No, it's, it's actually your fault, right? For putting your stuff there. So I think there's more going on than just price discussions and, and, but there is a bit of a repatriation in some instances of things coming back on cloud while cloud still continues to grow, right?

Um, so it's, it's a very interesting dynamic right now, How much of that do you think may be just a post covid phenomenon? Because, you know, during covid, the only game in town was to put things in the cloud. And maybe people are now looking at that and saying, you know, certain applications are more expensive in the club and they're on premise and they're just moving all that stuff back.

I'm sure COVID has a, has a play in it in the post covid world is is a little different, right? Um, and in some ways we're kind of, listen, it is is like a pendulum that goes back and forth over the years and, you know, we can go back in the, uh, in the two thousands where everything was outsourced, outsourced, outsourced, and then in the late two thousands it was insource, insource, insource, and then cloud, cloud, cloud, and now hybrid, right? And it is like this all the time.

There's always gonna be this pendulum moving back and forth. Um, and right now we're in this, I think in, in a hybrid pendulum, to be honest with you, where some things have to stay on premises because there's a long tail of Windows applications that isn't going away anytime soon. And, um, but there's also the desire to move to the cloud because of convenience.

And in some ca cases it's actually cheaper from a total cost of ownership perspective. And I, and I think part of the reason this is coming up now is maybe the proper analysis wasn't done on the total cost of ownership of running your own data center, the total cost of ownership, right? And now instead of having all of that spread out and opex and CapEx costs, I get one bill that has everything in it from that provider that I was probably paying anyway, but in, in different buckets.

So it looks a lot worse. True. So the analysis needs to be kinda put into some context.

Um, let me ask you this question though. Um, do you think that, um, what is the definition of cloud? And I'm asking this question because if I go watch, you know, Dell and hp, they'll tell me it's a cloud operating model and I'm dropping a cloud like workload in an on-premises environment and I'm paying for it with a, an OPEX expense.

And so if everything is a cloud, then maybe nothing's a cloud. And has the term kind of gone past its useful life? You know, I used to joke cloud AKA internet, But that's not exactly true.

It's funny you ask that. 'cause I ask that on interview questions. That's one of my interview questions I ask people.

'cause I like to have their understanding of what cloud is because it's, it's different depending on your point, point of view. If you ask somebody that came from the storage world, what the cloud is, they're going to give you a storage point of view of what they think the cloud is. If you ask somebody in the entertainment industry what the cloud is, they're gonna reference Netflix and Prime video, right?

Delivering content from the cloud. Uh, if you ask somebody that's in the, uh, infrastructure world, they're gonna talk about infrastructure as a service. And if you talk to somebody in the dev world, they're gonna talk about PAs platform as a service.

It actually means something different to everybody. It's a, it's a very generic term at these at this point. Um, and I do think it, it's, it's a matter of your point of view, but the reality is it's, it's, it is resources being delivered from the internet, right?

And the internet could be somebody, somebody, uh, else's data center or it could be your data center. I mean, if you wanna really simplify it, that's, that's really what's happening at a very simple level. And as the days of capital expenditures done for IT infrastructure, because I've talked to some folks and, you know, they're in the oil and gas industry, they're like, no, we got more cash than we know what to do with.

So they're not really excited about operating expenses. They kinda like capital budgets. And so does this weight shift over time, depending on how flush you are or what, what tips you one way or the other?

It it's the pendulum and I think it's individual to each company's financials. Some companies depend more heavily on opex, others on CapEx. It really is completely individual, uh, to every company.

And so the trend is, is is going to change all the time, just like the pendulum I talked about. So ultimately, what's your best advice to folks given this pendulum that keeps swinging? I mean, a lot of us are getting whiplash, but how do I wrap my arms around this?

I Think take the time to do a good analysis of total cost of ownership. Of course, your financial organization that BU is gonna have a good understanding of whether opex or CapEx is better for them. If they wanna depreciate and that's good for them, great.

If they don't want to, and it hurts their books or the way that they're reporting if they're a public company, the way they're reporting, yeah, every company's individual. So I, I would just argue that they have to take the time to figure out what are their primary goals, what's their north star, what's gonna get 'em there, what are the metrics they need to, to to work towards? And then you figure out is it a cop CapEx or, or an opex situation for you.

All right, folks. I guess I don't know why cloud prices are consistent anymore than I know why gasoline prices seem to be consistent from station to station, but it's a mystery. But I'm sure somebody will eventually look into it, but in the meantime it's yours to control.

So the more you know, the smarter you are, the less it's gonna cost. Hey Brad, thanks for being on the show. Thank you.

It's been my pleasure. I appreciate it. Thanks for having, Thanks.

Well that.